Use of Uninitialized Resource

Summary

In TensorFlow before 1.15.5, 2.0.0 through 2.0.3, 2.1.0 through 2.1.2, 2.2.0 through 2.2.1, and 2.3.0 through 2.3.1, under certain cases a saved model can trigger use of uninitialized values during code execution. This is caused by having tensor buffers be filled with the default value of the type but forgetting to default initialize the quantized floating point types in Eigen. This also affects tensorflow-cpu and tensorflow-gpu packages. It's fixed in versions 1.15.5, 2.0.4, 2.1.3, 2.2.2, 2.3.2, and 2.4.0.